Research Article

Research on Discriminative Skeleton-Based Action Recognition in Spatiotemporal Fusion and Human-Robot Interaction

Table 3

Comparison of CS and CV benchmarks with state-of-the-art schemes.

MethodsCS (%)CV (%)

STA-LSTM [6]73.481.2
ST-GCN [8]81.588.3
CNN-based [7]83.289.3
GCN-NAS [14]89.495.7
MS-AAGCN [10]90.096.2
2s-AGCN + PM-STFGCN (ours)91.996.5